Free image tools

Compare two images side by side and spot the differences

Load two pictures to inspect them with a before/after slider or a difference view. The tool reports each image's dimensions and the percentage of pixels that differ.

This tool needs JavaScript. It runs entirely in your browser; nothing is uploaded.

Private by design. Your files are processed by your own browser on your own device. They are not uploaded, stored or seen by Archive.London. Tool terms

How to use the Compare Two Images tool

  1. Choose the first image, then the second.
  2. Drag the before/after slider across the picture to compare the two.
  3. Switch to the difference view to see which areas have changed.
  4. Read the reported dimensions and the percentage of differing pixels.

To compare two images, open the Compare Two Images tool on Archive.London and load both files. Drag the before/after slider to inspect them against each other, or switch to the difference view to see exactly where they diverge. The tool reports the dimensions of each and the percentage of differing pixels. Nothing is uploaded.

Slider or difference view: which to use

The before/after slider is the better choice for judgements of quality. Drag it over fine detail, such as hair, fabric texture or lettering, to see what a compression setting, a sharpening pass or a colour correction has done. Your eye compares the same area in both versions with nothing else changing.

The difference view is for locating change. It shows where the two images diverge, so a retouched blemish, a cloned-out object or a replaced sky stands out even when the two pictures look alike at a glance. Areas that match show as no difference.

The percentage of differing pixels summarises the result as a number. It tells you how much of the frame changed, not how badly. A faint overall colour shift can alter nearly every pixel, while a significant local edit may touch a fraction of one per cent.

Why two copies of the same photo are rarely pixel-identical

A common surprise is a high difference figure between two files that look the same. Several ordinary processes cause it.

  • JPEG compression is lossy and works in 8 × 8 pixel blocks. Every re-save at a different quality nudges values across the whole image.
  • Resizing recalculates every pixel.
  • Colour management matters. Images are compared as the browser decodes them, so two files with different embedded profiles can differ in value even when they are meant to look alike.
  • Different dimensions mean the pixels do not line up one to one. For a meaningful percentage, compare images with the same pixel dimensions.

Lossless round trips are the exception. A PNG converted from another lossless file should show no differing pixels, which makes this a quick check that a conversion preserved the picture.

Comparison stays on your device

Both images are decoded and compared in your browser. Neither is uploaded, and nothing is stored after you close the page. That makes the tool suitable for work that cannot be sent to a third party: unreleased campaign images, retouching proofs awaiting approval, or evidence in a dispute over whether a published picture was altered.

There are practical limits. Each image is held in memory as uncompressed pixels, so comparing two very large scans needs a computer with plenty of memory, and files over about 100 megapixels may fail on a phone. The tool reports and displays differences. It does not save an annotated result as a new file, so take a screenshot if you need a record of what you saw.

Comparing scans, duplicates and published versions

Comparison is everyday work in a film archive. A transparency may have been duplicated, a frame may survive as both an original and a print, and the version that ran in a magazine may have been cropped or colour-corrected from the film. Setting two scans against each other shows which is sharper, which carries more dust and scratches, and whether two similar frames from a sequence are in fact the same exposure.

The same technique lets picture researchers check a supplied file against a published reference. The Robert Harper Photograph Library holds this kind of layered material, including transparencies, negatives and contact sheets, described under analogue holdings.

Frequently asked questions

How do I find the differences between two pictures?

Load both images and switch to the difference view. Areas where the pictures match show no difference, and areas that have changed stand out, so edits are easy to locate. The tool also gives the percentage of pixels that differ. For subtle changes in quality, the before/after slider is more useful because you can examine the same detail in each version.

Why does the tool say most pixels are different when the images look the same?

Small numerical changes are invisible to the eye but still count as differences. Saving a JPEG again, resizing, or converting between colour profiles shifts values slightly across the entire picture. The percentage measures how many pixels changed, not how visible the change is. Use the slider to judge whether the difference matters for your purpose.

Can I compare images of different sizes?

You can load them, and the tool reports the dimensions of each so a mismatch is obvious. A pixel-by-pixel percentage is only meaningful when both images have the same pixel dimensions, because otherwise the pixels do not correspond. Resize or crop one to match the other first if you need an accurate figure.

Can this prove that a photo has been edited?

It can show that two files differ and where. It cannot tell you which is the original, when a change was made or who made it, and it needs two versions to work. Ordinary recompression also produces differences. For questions of authenticity, combine a comparison with other evidence such as metadata, file hashes and, where they exist, Content Credentials.

Are my images uploaded to a server?

No. Both files are opened in your browser and compared on your own device. Nothing is transmitted or kept, and the page continues to work offline once it has loaded. This means unpublished or confidential pictures can be compared without sending them to anyone.

Licensing the archive for AI and data projects

95,487 photographs from one photographer and one rights holder. Read the dataset card, search the 503-folder inventory, discuss a sample set under NDA, or arrange a first look as scanning proceeds.

See AI and data licensingWrite to the archive